Re: [R-sig-phylo] PGLS transformations
Dear Sergio, I do not use R very much, relying instead on our Matlab programs that accompany Lavin et al. (2008): Lavin S.R., W.H. Karasov, A.R. Ives, K.M. Middleton, and T. Garland Jr. 2008. Morphometrics of the avian small intestine compared with that of nonflying mammals: a phylogenetic approach. Physiological and Biochemical Zoology 81:526–550. I would strongly suggest that you read the appendix to this paper regarding methods, their history, and terminology. Note that a GLS model does not inherently use any kind of branch length transformation. It uses the tree as inputted. It is mathematically equivalent to Felsenstein's (1985) phylogenetically independent contrasts. It implicitly assumes character evolution (or residual character evolution) akin to Brownian motion. Once you start estimating some branch lengths transformation along with regression parameters, you have a different beast. See the appendix noted above and, for the original source of this sort of model, Grafen (1989). Cheers, Ted Theodore Garland, Jr., Professor Department of Biology University of California, Riverside Riverside, CA 92521 Office Phone: (951) 827-3524 Facsimile: (951) 827-4286 (not confidential) Email: tgarl...@ucr.edu http://www.biology.ucr.edu/people/faculty/Garland.html http://scholar.google.com/citations?hl=en&user=iSSbrhwJ Director, UCR Institute for the Development of Educational Applications Editor in Chief, Physiological and Biochemical Zoology Fail Lab: Episode One http://testtube.com/faillab/zoochosis-episode-one-evolution http://www.youtube.com/watch?v=c0msBWyTzU0 From: R-sig-phylo [r-sig-phylo-boun...@r-project.org] on behalf of Sergio Ferreira Cardoso [sff.card...@campus.fct.unl.pt] Sent: Sunday, April 12, 2015 12:47 PM To: r-sig-phylo@r-project.org Subject: [R-sig-phylo] PGLS transformations Hi everyone, I'm relatively new in phylogenetic comparative methods. I'm a little confused about branch length transformations. I'm using a tree with divergence time (My) as branch lengths. When I use corPagel, corGrafen or corMartins in R, the branch lengths, are the branch lengths automatically transformed? e.g., gr.mammals<-corGrafen(1,phylo,fixed=F); fit<-gls(FCL~logBodymass,correlation=gr.mammals,data=df,method="ML"). My question may sound a bit nonsense but I've seen in some papers (e.g., Spoor, F., Garland, T., Krovitz, G., Ryan, T. M., Silcox, M. T., & Walker, A. (2007). The primate semicircular canal system and locomotion. *Proceedings of the National Academy of Sciences*, *104*(26), 10808-10812.) the indication that a PGLS was made without branch transformation, but no reference is made to the model (maybe it's corBrownian). Thank you very much. Best regards, Sérgio. -- Com os melhores cumprimentos, Sérgio Ferreira Cardoso. Best regards, Sérgio Ferreira Cardoso MSc. Paleontology candidate Departamento de Ciências da Terra - FCT /Universidade Nova de Lisboa Geociências - Universidade de Évora Lisboa, Portugal [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/ ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
[R-sig-phylo] PGLS - branch lengths 1
Dear all, I'm trying to perform a PGLS with arbitrary branch lengths (I used branch lengths = 1). The tree is non-ultrametric. Here is a result with Pagel's Lambda: vf<-diag(vcv(tree)) fit<-gls(FCLrelative~LogBodymass,correlation=Pagel,data=DF,method="ML",weights=varFixed(-vf)) summary(fit) Model: FCLrelative ~ LogBodymass Data: DF AIC BIClogLik 42.28777 49.77257 -17.14388 Correlation Structure: corPagel Formula: ~1 Parameter estimate(s): lambda 1.039835 Variance function: Structure: fixed weights Formula: ~vf Coefficients: Value Std.Errort-value p-value (Intercept)-0.17662926 0.26166738 -0.6750144 0.5030 LogBodymass 0.058657370.057661711.0172672 0.3143 Correlation: (Intr) LogBodymass -0.651 Standardized residuals: Min Q1 Med Q3 Max -2.36346640 -0.21848447 0.08326376 0.46156642 1.15440168 Residual standard error: 0.2388999 Degrees of freedom: 48 total; 46 residual As you see my lambda is >1, what shouldn't happen. Is this normal when we set all branch lengths to an arbitrary value? Thank you very much in advance. Best regards, Sérgio. -- Com os melhores cumprimentos, Sérgio Ferreira Cardoso. Best regards, Sérgio Ferreira Cardoso MSc. Paleontology candidate Departamento de Ciências da Terra - FCT /Universidade Nova de Lisboa Geociências - Universidade de Évora Lisboa, Portugal [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
[R-sig-phylo] PGLS transformations
Hi everyone, I'm relatively new in phylogenetic comparative methods. I'm a little confused about branch length transformations. I'm using a tree with divergence time (My) as branch lengths. When I use corPagel, corGrafen or corMartins in R, the branch lengths, are the branch lengths automatically transformed? e.g., gr.mammals<-corGrafen(1,phylo,fixed=F); fit<-gls(FCL~logBodymass,correlation=gr.mammals,data=df,method="ML"). My question may sound a bit nonsense but I've seen in some papers (e.g., Spoor, F., Garland, T., Krovitz, G., Ryan, T. M., Silcox, M. T., & Walker, A. (2007). The primate semicircular canal system and locomotion. *Proceedings of the National Academy of Sciences*, *104*(26), 10808-10812.) the indication that a PGLS was made without branch transformation, but no reference is made to the model (maybe it's corBrownian). Thank you very much. Best regards, Sérgio. -- Com os melhores cumprimentos, Sérgio Ferreira Cardoso. Best regards, Sérgio Ferreira Cardoso MSc. Paleontology candidate Departamento de Ciências da Terra - FCT /Universidade Nova de Lisboa Geociências - Universidade de Évora Lisboa, Portugal [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
Re: [R-sig-phylo] Unsual values of alpha in OU models
Thank you all for your responses, they have been of great use. First, I have not rescaled the tree, thus alpha values are really rare and they indicate a rate adaption very fast. For the selection of models I have used AICc and then I have analysed the reliability of the parameters with a bootstrap (nboot = 100) as you have advised me. The results show that alpha and sigma.sq parameters have wide confidence intervals (more than 1) on all models, while theta presents narrow intervals in two models (OUM and OUMV) but wide in the two most parsimonious models (OUMA and OUMVA). I thought take best estimated theta values but I have noticed that there is a change in level 3 of my discrete variable: in OUM-OUMV optimum is 0 (optimal value of level 2) while in OUMA-OUMVA optimum is 2 (optimum value of level 1). It seems clear that there are two optima but I don't know that occur with the level 3 of discrete variable, which is an intermediate situation between levels 1 and 2. Thanks for everything Diego Salazar 2015-03-28 14:25 GMT+00:00 Aaron King : > I agree with Brian that bootstrapping or likelihood profiles are a good > way to determine confidence intervals on your parameters. But I disagree > with his suggestion that ability to estimate parameters is a proper > component of model selection. It's important to realize that the width of > confidence regions is an indication of the information content of the data > relative to those parameters. The fact that two parameter values have > equal or very similar likelihoods means that the data are agnostic as to > their relative explanatory value. To put it another way, if you are asking > a question that hinges on the value of a parameter and the confidence > interval for that parameter is so wide that your question goes unanswered, > this is because the data do not contain an answer to your question. > Therefore, the only way to get such an answer is to get better data. One > is, of course, always free to make additional or different assumptions > (such as a different model as Brian suggests or imposing a prior), but do > not make the mistake of confusing assumptions with information. This is > what you would be doing if you were, for example, to reject a model with a > high likelihood but poorly identified parameters in favor of one with a low > likelihood but precise estimates. > > -- > Aaron A. King, Ph.D. > Ecology & Evolutionary Biology > Mathematics > Center for the Study of Complex Systems > University of Michigan > GPG Public Key: 0x15780975 > -- Diego Francisco Salazar Tortosa Ph student Department of Ecology University of Granada Av. Fuente Nueva s/n 18071 Granada Telefono: +34 958241000 ext 20007 Movil: +34 634851132 email: dsala...@ugr.es dftort...@gmail.com --- "Este mensaje se dirige exclusivamente a su destinatario y puede contener información privilegiada o confidencial. Si no es Ud. el destinatario indicado, queda notificado de que la utilización, divulgación o copia sin autorización está prohibida en virtud de la legislación vigente. Si ha recibido este mensaje por error, se ruega lo comunique inmediatamente por esta misma vía y proceda a su destrucción. This message is intended exclusively for its addressee and may contain information that is CONFIDENTIAL and protected by professional privilege. If you are not the intended recipient you are hereby notified that any dissemination, copy or disclosure of this communication is strictly prohibited by law. If this message has been received in error, please immediately notify us via e-mail and delete it". -- [[alternative HTML version deleted]] ___ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/